Cm

FJMa

b

c

d

a

ARR1A

KCCBIS

1

OigbacotMeeL

rec

0d

Journal of Neuroscience Methods 195 (2011) 135–142

Contents lists available at ScienceDirect

Journal of Neuroscience Methods

journa l homepage: www.e lsev ier .com/ locate / jneumeth

haracterisation of carbon paste electrodes for real-time amperometriconitoring of brain tissue oxygen

iachra B. Bolgera, Stephen B. McHughb, Rachel Bennetta, Jennifer Li c, Keita Ishiwaria,c,ennifer Francoisc, Michael W. Conwayc, Gary Gilmourc, David M. Bannermanb, Marianne Fillenzd,

ark Tricklebankc, John P. Lowrya,∗

Department of Chemistry, National University of Ireland Maynooth, Maynooth, Co. Kildare, IrelandDepartment of Experimental Psychology, University of Oxford, South Parks Road, Oxford OX1 3UD, UKCentre for Cognitive Neuroscience, Eli Lilly and Co. Limited, Windlesham, Surrey, UKDepartment of Physiology, Anatomy and Genetics, University of Oxford, Parks Road, Oxford OX1 3PT, UK

r t i c l e i n f o

rticle history:eceived 11 August 2010eceived in revised form1 November 2010ccepted 21 November 2010

a b s t r a c t

Tissue O2 can be monitored using a variety of electrochemical techniques and electrodes. In vitro andin vivo characterisation studies for O2 reduction at carbon paste electrodes (CPEs) using constant potentialamperometry (CPA) are presented. Cyclic voltammetry indicated that an applied potential of −650 mV isrequired for O2 reduction at CPEs. High sensitivity (−1.49 ± 0.01 nA/�M), low detection limit (ca. 0.1 �M)and good linear response characteristics (R2 > 0.99) were observed in calibration experiments performed

eywords:arbon paste electrodesonstant potential amperometryrain tissue oxygen

n vivo electrochemistry

at this potential. There was also no effect of pH, temperature, and ion changes, and no dependence uponflow/fluid convection (stirring). Several compounds (e.g. dopamine and its metabolites) present in brainextracellular fluid were tested at physiological concentrations and shown not to interfere with the CPAO2 signal. In vivo experiments confirmed a sub-second response time observed in vitro and demonstratedlong-term stability extending over twelve weeks, with minimal O2 consumption (ca. 1 nmol/h). These

operuous

ensors results indicate that CPEsbe used reliably to contin

. Introduction

Brain cells are critically dependent on a continuous supply of2 for their normal energy metabolism with the brain consum-

ng approximately 20% of the total O2 used by the body at anyiven time. The tissue concentration is determined by the balanceetween blood supply and local utilization. A decrease in supply, orlarge increase in consumption without adequate compensation,

an seriously compromise brain function due to the low reservef dissolved O2 present in the tissue; the distribution of concen-rations reported ranges from 40 �M to 80 �M (Nair et al., 1987;

cCreery et al., 1990; Kayama et al., 1991; Murr et al., 1994; Zaunert al., 1995) depending on the depth of measurement (Baumgärtlt al., 1989), and the heterogeneity of the tissue (Murr et al., 1994;üebbers and Baumgäertl, 1997).

Oxygen levels in the brain can be measured in several ways: indi-ectly, using non-invasive near-infrared spectroscopy (Matsumotot al., 1996; Rasmussen et al., 2007); globally, using fibre-opticatheters to monitor jugular venous O2 saturation (SjO2) (Coplin

∗ Corresponding author. Tel.: +353 1 708 3770/4639; fax: +353 1 708 3815.E-mail address: john.lowry@nuim.ie (J.P. Lowry).

165-0270/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.jneumeth.2010.11.013

ating amperometrically at a constant potential of −650 mV (vs. SCE) canly monitor brain extracellular tissue O2.

© 2010 Elsevier B.V. All rights reserved.

et al., 1998; Gopinath et al., 1999; Howard et al., 1999); and locally,using Clark-type electrode technology to directly monitor O2 par-tial pressure (pO2) (Clark et al., 1958; Thompson et al., 2003;Piilgaard and Lauritzen, 2009). The latter operates by measuringthe electrochemical reduction of O2 and can be used in freely mov-ing animals. It also offers significant spatial (∼10 �m) and temporal(millisecond) advantages compared to other techniques.

Since the pioneering ‘brain polarography’ research carried outby Clark and colleagues (Clark et al., 1953; Thompson et al., 2003)over 50 years ago a wide variety of electrodes (sensors) havebeen used. These can essentially be divided into two main groups:noble metal electrodes, such as Pt (Clark et al., 1958; Travis andClark, 1965; Thompson et al., 2003; Offenhauser et al., 2005) andAu (Cooper, 1963; Holmström et al., 1998; El-Deab and Ohsaka,2003); and carbon-based electrodes, such as glassy carbon (Clarkand Clark, 1964), carbon fibre (Zimmerman and Wightman, 1991;Zimmerman et al., 1992; Venton et al., 2003), carbon epoxy (Bazzuet al., 2009), and carbon paste (CPE) (Lowry et al., 1996, 1997; Bolger

and Lowry, 2005). While carbon electrodes tend to be more labourintensive in terms of their construction, they have the advantagethat they are less prone to surface poisoning and as such do notrequire the use of protecting membranes which are a character-istic of metal-based O2 electrodes. Of the three types used, glassy

1 oscience Methods 195 (2011) 135–142

cwcdvti(gaa(ddsd

tiu(toitlChsiowssiW

2

2

NMLa((rh(H(obg

i0wa(b1dw

Fig. 1. Typical in vitro current–time response for an O2 calibration (0–1200 �M; N2,air and O2-saturation) at a carbon paste electrode (CPE) carried out using constant

36 F.B. Bolger et al. / Journal of Neur

arbon can be restricted by its large size (>1 mm) (Jin et al., 2005),hile with carbon fibre electrodes their small dimension (typi-

ally 10 �m) means that the concentration of O2 observed can varyepending on the orientation of the electrode relative to the bloodessels and metabolically active sites, and on the depth of penetra-ion into the tissue (Baumgärtl et al., 1989). The latter is particularlymportant when using cylinder electrodes. Since the dimensiontypically 100–200 �m) (Justice, 1987) of carbon epoxy and CPEs isreater than the scale of a capillary zone (ca. 70 �m) (Silver, 1965),n average tissue O2 level is detected, but this must be balancedgainst the increased tissue damaged caused by the larger electrodeDuff and O’Neill, 1994; Khan and Michael, 2003). While micro-ialysis probes of similar dimensions have been reported to alteropamine levels (Khan and Michael, 2003) there is no evidence touggest that this is the case for an electrochemically detected freelyiffusing gaseous species such as O2.

We have previously reported preliminary data demonstratinghat CPEs can be used to cathodically monitor real-time changesn brain tissue O2 during neuronal activation (physiological stim-lation) in freely moving rodents using both differential pulseLowry et al., 1996; Bolger and Lowry, 2005) and constant poten-ial (Lowry et al., 1997) amperometry (DPA and CPA). The increasesbserved with DPA were also found to correlate with increasesn regional cerebral blood flow measured using the H2 clearanceechnique (Lowry et al., 1997). While we have subsequently pub-ished detailed characterisation results for DPA O2 monitoring atPEs (Lowry et al., 1996) we now present the results of compre-ensive characterisation studies for CPA, a technique which haseveral advantages compared to pulsed methods, including simplenstrumentation and experimental design (i.e. it does not requireptimized pulse sequences), and continuous real-time recordingith high sensitivity and low background (baseline) noise. Such

tudies are important in establishing the properties confirminguitability for in vivo monitoring and include testing sensitiv-ty, selectivity, consumption/depletion and stability (Phillips and

ightman, 2003).

. Materials and methods

.1. Chemicals and solutions

The NaCl (SigmaUltra), NaH2PO4 (Sigma, A.C.S. reagent),aOH (SigmaUltra), KCl (SigmaUltra), CaCl2 (SigmaUltra), andgCl2 (SigmaUltra) were used as supplied (Sigma–Aldrich Ireland

td). Compounds used in the interference study were: l-scorbic acid (AA; A.C.S. reagent, Sigma), dehydroascorbic acidDHAA; Aldrich), uric acid (UA; sodium salt, Sigma), glutathioneoxidised disodium salt, Aldrich), dopamine (DA; hydrochlo-ide, Sigma), 3,4-dihydroxyphenylacetic acid (DOPAC; Sigma),omovanilic acid (HVA; Fluka Biochemika), 5-hydroxytryptamine5-HT; hydrochloride, Sigma), 5-hydroxyindole-3-acetic acid (5-IAA; Fluka Biochemika), l-tryptophan (99%, Aldrich), l-cysteine

>98%, Sigma), l-tyrosine (99%, Aldrich). Stock standard solutionsf all compounds were prepared from the supplied chemicals at theeginning of each experiment to avoid problems associated withradual decomposition.

Unless otherwise stated in vitro experiments were carried outn phosphate buffer saline (PBS) solution, pH 7.4 (0.15 M NaCl,.04 M NaH2PO4 and 0.04 M NaOH). In pH studies the buffer pHas adjusted to between 6.5 and 8.0 using solutions of NaH2PO4

nd NaOH. Experiments investigating the effects of ion changes

Ca2+ and Mg2+) on O2 sensitivity were performed in artificial cere-rospinal fluid (aCSF): 147 mM NaCl; 4 mM KCl; 1.2 mM CaCl2; andmM MgCl2 (Deboer et al., 1990). All solutions were prepared usingeoxygenated doubly distilled deionised water and stored at 4 ◦Chen not in use.

potential amperometry (CPA) at −650 mV (vs. SCE) in PBS, pH 7.4. Inset: typicalexample of the effect of changing the O2 concentration (bubbling air) on the CPEresponse characteristics for O2 reduction. CPE background currents subtracted.

2.2. Working electrode preparation

Carbon paste was prepared by thoroughly mixing 0.71 g ofgraphite powder (1–2 �m, Aldrich) with 250 �L of silicone oil (hightemperature, Aldrich) (O’Neill et al., 1982). CPEs (8 T, 200-�m barediameter, 256-�m coated diameter) were either made in-housefrom Teflon-coated silver wire (Advent Research Materials, Suffolk,UK) as reported previously (Lowry et al., 1997) or supplied by BlueBox Sensors Ltd (Dublin, Ireland). When not in use all electrodeswere stored in PBS at 4 ◦C.

2.3. Instrumentation and software

All electrochemical techniques (constant potential amperom-etry (CPA) and cyclic voltammetry) were carried out using alow-noise potentiostat (Biostat IV, ACM Instruments, Cumbria, UK).Data acquisition was performed with a notebook PC, a PowerLabinterface system (ADInstruments Ltd., Oxford, UK) and LabChartfor Windows software (ADInstruments Ltd.).

All data are presented as mean ± standard error (SEM), withn = number of electrodes. All analysis was performed usingMicrosoft Excel 2007 and the commercial packages Prism (version5.01) and InStat (GraphPad Software Inc., CA, USA). The statisticalsignificance of differences observed was calculated using Student’st-tests (two-tailed paired or unpaired observations where appro-priate) or one-way ANOVA (Kruskal–Wallis test with Dunn’s posttest). Values of P < 0.05 were considered to indicate statistical sig-nificance.

2.4. Experiments in vitro

Experiments in vitro were performed in a standard three-electrode glass electrochemical cell containing 15 mL PBS at roomtemperature unless otherwise stated. A saturated calomel elec-trode (SCE) was used as the reference electrode, and a Pt wireserved as the auxiliary electrode. CPEs were allowed settle underthe influence of the applied CPA potential (−650 mV or 0 mV vs.SCE) until the non-faradaic current had reached a stable baselinelevel—typically 30 min.

To attain effective deaeration, the PBS solution was vigorouslypurged with O2-free N2 (BOC Ireland, average O2 content 2 ppm,maximum O2 content 5 ppm) for at least 30 min before recordingbegan. In calibration experiments involving 0–1200 �M solution O2

either N2, atmospheric air (from a RENA air pump) or pure O2 (com-pressed gas) was bubbled through the PBS for a similar period andthe appropriate gaseous atmosphere then maintained over the cellsolution during quiescent recording (see Fig. 1A). The concentra-tions of solution O2 were taken as 0 �M (N2-saturated), 240 �M

oscien

((

+wfcLtctabpItnsPr

2

acBllMaD±(waptwtaaaul(

2

dwatpawtat

trsco

(quiescent), −628 ± 73 (1 Hz); P = 0.83, n = 8) and represents anincrease in current (3.8 ± 2.4%) similar to that observed at Pt micro-electrodes (ca. 2–3%) (Sharan et al., 2008). While this data is morerelevant for in vitro use in brain slice or cell culture experiments

F.B. Bolger et al. / Journal of Neur

air-saturated (Foster et al., 1993)) and 1200 �M (O2-saturatedBourdillon et al., 1982)) respectively.

In calibrations involving 0–240 �M O2 known volumes (+200,204, +208, +212, and +216 �L) of a saturated (100%) O2 solutionere added to 10 mL of N2-purged PBS. Mixing in these, and inter-

erence experiments, was achieved by placing the electrochemicalell on a magnetic stirrer (IKA MST Mini Magnetic stirrer, Lennoxaboratory Supplies Ltd, Dublin, Ireland) and agitating (ca. 10 Hz)he cell solution using a stirring bar (20 mm × 5 mm diameter) fora. 5 sec following each injection aliquot. In studies investigatinghe effects of fluid convection stirring was carried out 1 Hz undern air-saturated atmosphere. Calibrations at 37 ◦C were performedy placing the cell on a temperature regulated magnetic stirrer/hotlate (IKA MST Basic C, Lennox Laboratory Supplies Ltd, Dublin,

reland). The solution temperature was controlled using a TC 1emperature controller (IKA) which was placed in the solution asear as possible to the sensor. Unless otherwise stated a N2 atmo-phere was maintained over all solutions throughout recording.ost in vivo calibrations were carried out in the 0–1200 �M O2ange.

.5. Surgery

Male Sprague–Dawley or Wistar rats weighing 200–350 g werenesthetised and implanted with CPEs (Day 0), following proto-ols similar to those previously described (Lowry and Fillenz, 2001;olger and Lowry, 2005). The level of anesthesia was checked regu-

arly (pedal withdrawal reflex). Typical coordinates, with the skullevelled between bregma and lambda, were: A/P +1.0 from bregma,

/L ±2.5, and D/V −5.0 (striatum); A/P −3.6 from bregma, M/L ±2.2,nd D/V −3.2 (hippocampus); A/P +2.7 from bregma, M/L ±1.2, and/V −3.8 (medial prefrontal cortex); A/P +1.9 from bregma, M/L0.8, and D/V −6.9 (nucleus accumbens). A reference electrode

8 T Ag wire, 200-�m bare diameter; Advent Research Materials)as placed in the cortex and an auxiliary electrode (8 T Ag wire)

ttached to one of the support screws (see below). The referenceotential provided by the Ag wire in brain tissue is very similaro that of the SCE (O’Neill, 1993). Unless otherwise stated animalsere allowed recover from surgery with the electrodes fixed to

he skull using dental screws and dental acrylate. All animals weressessed for good health according to published guidelines (Mortonnd Griffiths, 1985) immediately after recovery from anesthesiand at the beginning of each day. All procedures were performednder license in accordance with the European Communities Regu-

ations 2002 (Irish Statutory Instrument 566/2002 and UK AnimalsScientific Procedures) Act 1986).

.6. Experiments in vivo

Rats were housed in a windowless room under a 12 h light, 12 hark cycle, lights coming on at 8 am, with free access to water. Foodas available ad libitum. All experiments were carried out with the

nimal in its home bowl. Implanted electrodes were connected tohe potentiostat after the 24 h recuperation period, through a six-in Teflon socket (MS363, Plastics One, Roanoke, VA, USA), andflexible screened six core cable (363-363 6TCM, Plastics One)hich was mounted through a swivel (SL6C, Plastics One) above

he rat’s head. This arrangement allowed free movement of thenimal which remained continuously connected to the instrumen-ation.

Mild hypoxia and hyperoxia were produced by the administra-

ion of N2/air and O2/air mixtures; plastic tubing, connected to theespective gas cylinder (BOC), was held ca. 2–3 cm from the animal’snout for either 30 (N2/air) or 60 s (O2/air) periods. A flow rate ofa. 150 mL/min was used. This procedure resulted in the inhalationf an air/gas mixture.

ce Methods 195 (2011) 135–142 137

3. Results and discussion

3.1. Oxygen reduction

Electrochemical reduction of O2 at carbon electrodes is a two-electron processes producing H2O2:

O2 + 2H+ + 2e− → H2O2

H2O2 + 2H+ + 2e− → 2H2O

Since the direct reduction (Martel and Kuhn, 2000) (and oxi-dation (Taylor and Humffray, 1975; Zimmerman and Wightman,1991)) of H2O2 is severely inhibited at carbon electrode surfacesthe rate-limiting step is the initial one-electron step followed byprotonation of the superoxide ion and further reduction (Taylorand Humffray, 1975). We have previously determined the positionof O2 reduction on the voltage axis of CPEs using cyclic voltammetry(Lowry et al., 1996). Characterisation voltammograms recorded at100 mV/s in N2 (background) and air-saturated (21% O2) PBS solu-tions indicate that a potential of −650 mV (vs. SCE) is appropriatefor CPA O2 detection; this potential is in the mass-transport limitedregion after the peak potential for O2 reduction (ca. −500 mV). Inorder to determine the O2 sensitivity of CPEs operating at −650 mVcalibrations were performed over a 0–1200 �M O2 range (N2, airand O2-saturated PBS). It takes approximately 5 min for the buffersolution to reach the new level of gaseous saturation (Fig. 1). Thedecrease in current during the quiescent periods is a result of theremoval of forced convection due to the bubbling associated withthe introduction of the atmospheric air (air pump) and pure O2(compressed gas). The response was linear (R2 = 0.999, n = 8) witha slope of −1.49 ± 0.01 nA/�M, n = 8. There is an instantaneouschange in the current upon increasing the O2 concentration viabubbling with the response time <1 s (inset Fig. 1). This is com-parable to the O2 response characteristics reported for Clark-typenoble metal electrodes (Piilgaard and Lauritzen, 2009), but withthe added advantage of long-term in vivo stability (see Section 3.6).Calibrations performed over a range of low, physiologically rele-vant, concentrations (0–125 �M O2) produced a similar sensitivity(−1.09 ± 0.03 nA/�M, R2 = 0.998, n = 4) and subsecond responsetime (Fig. 2A and B). The calculated limit of detection (3 × SD ofthe background noise level) of 0.09 �M is smaller than that foundfor DPA1 (8 �M) reflecting the low background currents that distin-guish CPA from DPA (Lowry et al., 1997; Bolger and Lowry, 2005).When the same calibrations were repeated at 0 mV (vs. SCE) therewas no change in the background signal in response to increasingconcentrations of O2 (Fig. 2C).

Oxygen electrode signals are generally tested for their depen-dence on flow which is usually achieved in vitro by examining thesignal sensitivity to fluid convection (Schneiderman and Goldstick,1978; Gotoh et al., 1961). The O2 reduction current at CPEs was thusmonitored in stirred and unstirred solutions with convection pro-duced using a magnetic stirrer (Sharan et al., 2008). While a smallsignal increase in the air-saturated (240 �M O2) quiescent responsewas observed with stirring this was not significant (−607 ± 70 nA

1 For DPA O2 reduction, two equally sized cathodic pulses are applied, the firstfrom a resting potential at −150 mV to −350 mV that corresponds to the foot ofthe reduction wave for O2 at treated (tissue modified) CPEs, and the second from−350 mV to −550 mV that corresponds to the peak of the reduction wave. The dif-ference in the current (�I) sampled during these respective pulse pairs is calculated,and changes in �I used as a measure of changes in O2.

138 F.B. Bolger et al. / Journal of Neuroscien

Fig. 2. (A) Current–time responses for CPE O2 reduction (n = 4) at physiologi-cally relevant concentrations carried out in PBS, pH 7.4, at −650 mV vs. SCE. (B)Amperometric calibration plot for the quiescent steady-state data shown in (A).(C) Current–time responses for CPE O reduction (n = 4) at 0 mV (vs. SCE). Arrowsit(

widvsaoeba

tN

Nicholson, 1995) and a reported tissue sampling area of 15 mm2

2

ndicate injections (+200, +204, +208, +212, and +216 �L) of a saturated O2 solu-ion (1.2 mM) yielding concentrations of 25, 50, 75, 100 and 125 �M O2. The hashedgray) lines in (A) and (C) represent the SEM.

e would, in reality, not expect to see even such small changesn vivo as it has been reported that O2 electrodes in brain tissueo not show a dependence upon flow (Cooper, 1963). While con-ection results in a reduced diffusion layer thickness at the sensorurface (Bard and Faulkner, 1980a; Amatore et al., 2000), and hencen increase in current, Cooper found that the “self stirring” actionf blood movement through the tissue causes this diffusion layerffect to be negligible in vivo (Cooper, 1963). This was explainedy the fact that the O concentration varies in accordance with the

2cceleration and not the velocity of the blood.

Confirmation that CPEs operating at −650 mV respond rapidlyo changes in brain tissue O2 in vivo was obtained by administering2 and O2 gas (Fig. 3A): 1-min of an O2/air mixture resulted in an

ce Methods 195 (2011) 135–142

increase of 50.1 ± 5.9 nA (n = 4) in the signal from baseline while 30-s of N2/air produced a decrease of 16.2 ± 4.3 nA (n = 4). Changes inboth cases were immediate. On cessation of inhalation of N2/air thesignals quickly returned to baseline levels indicating a rapid returnto normoxic conditions. For O2/air the return was slower and thesignals had not reached baseline during the period of recording.When the same experiments were repeated at an applied potentialof 0 mV (Fig. 3B) no change in signal was observed, supporting thein vitro calibration results obtained at 0 mV, and confirming that anapplied potential of −650 mV is required for O2 reduction at CPEs.Such data is in agreement with previously reported O2 characteri-sation data from in vivo experiments performed using DPA in freelymoving animals (Lowry et al., 1996, 1997; Bolger and Lowry, 2005):periods of mild hypoxia and hyperoxia produced rapid (subsecond)decreases and increases in the CPE signal; vasodilators such as thecarbonic anhydrase inhibitor acetazolamide (Diamox) increasedthe signal; while neuronal activation (tail pinch and stimulatedgrooming) produced similar increases in both regional cerebralblood flow (rCBF) and O2 indicating that CPE O2 currents pro-vide an index of increases in rCBF when such increases exceed O2utilization. The two techniques were found to have very similarsensitivities post implantation (Lowry et al., 1998), and in prelim-inary in vivo experiments involving tail pinch (Lowry et al., 1997)and insulin administration (Lowry et al., 1998) both methods pro-duced similar response characteristics. The advantage of CPA overDPA is that it has much lower background (capacitance) currents invivo (−49 nA vs. −748 nA) (Lowry et al., 1998), but is limited to thedetection of a single analyte (see Section 2.4), whereas DPA, on theother hand, can be used to simultaneously measure several species(e.g. O2 and ascorbic acid) (Lowry et al., 1996).

3.2. Oxygen consumption

As already outlined, in vivo voltammetry involves the detectionof substances in the extracellular fluid (ECF) using electrochemistrywith implanted microelectrodes (e.g. amperometric sensors andbiosensors) (Lowry and O’Neill, 2005); by implanting the electrodein a specific brain region, applying a suitable potential profile andrecording the resulting Faradaic current, changes in the concentra-tion can be monitored in real-time. However, such concentrationshave the potential to be altered by the voltammetric measure-ment itself through depletion in the vicinity of the electrode.While this is certainly the case for brain microdialysis (Sam andJustice, 1996; Borland et al., 2005), such effects tend to be small forvoltammetry and can be minimised for example with appropriateselection of sampling intervals and other parameters (Ewing et al.,1981).

The consumption of O2 in vivo by the CPE polarised at −650 mVwas thus determined using the following rate equation (Bard andFaulkner, 1980b): v = i/nFA, where v is the rate of the electrodereaction, i is the baseline current, n is the number of electronsper molecule of O2 reduced, F is the Faraday constant, and A isthe area of the sensor. Taking −60 nA (nC/s) as a typical baselinecurrent, 2 for the value of n (O2 reduction at a carbon surface),9.65 × 104 C/mol for the Faraday constant, and 3.14 × 10−4 cm2 forthe area of the sensor, yields a value of 0.001 �M O2 reduced/cm2/s.This corresponds to a consumption rate of ca. 1.13 nmol/h whichis comparable to the calculated value (ca. 0.92 nmol/h) for noblemetal electrodes of similar dimension. Assuming that the CPE sam-ples from a fluid pool of 50 �m radius at the electrode tip (Rice and

(Ma and Wu, 2008) this corresponds to a localized utilization(depletion) rate of ca. 12.5 nmol/g/min which is small relative toreported values 3.9 �mol/g/min (Madsen et al., 1998) for the cere-bral metabolic rate of O2 consumption (CMRO2).

F.B. Bolger et al. / Journal of Neuroscience Methods 195 (2011) 135–142 139

F 100%C the mr nly fo

cRmm

3

tf1dwrtCps(o

iaf1aH

ig. 3. Typical raw data (current–time) examples of the effects of mild hyperoxia (PEs (n = 4) at −650 mV (A) and 0 mV (B). Data from bilaterally implanted CPEs inepresent periods of gas administration. Hashed gray lines (shown above the data o

While electrodes with smaller dimensions will have smalleronsumption rates (e.g. 4 × 10−4–5 × 10−3 nmol/h for 10-�m Pt;evsbech, 1989; Piilgaard and Lauritzen, 2009) the concentrationsonitored can vary depending on proximity to blood vessels andetabolically active sites.

.3. Effect of temperature and pH

Classical membrane covered O2 sensors tend to have significantemperature dependence; the signal typically increasing by 1–6%or a rise of 1 ◦C (Hitchman, 1978; Jeroschewski and zur Linden,997). This is primarily due to the effects of temperature on theiffusion coefficient and the solubility of the gas in the membraneith temperature. As our in vitro experiments are routinely car-

ied out at room temperature (ca. 22.5 ± 0.2 ◦C), we thus examinedhe effect of increasing temperature on the CPE O2 sensitivity.alibrations (n = 4) in the range 0–240 �M O2 performed at thehysiological temperature of 37 ◦C (−1.93 ± 0.39 nA/�M) were notignificantly different (P = 0.95) from those at room temperature−1.90 ± 0.28 nA/�M). This is likely the result of the absence of anuter membrane on the CPE.

Changes in pH may occur during physiological experimentsn vivo (Zimmerman and Wightman, 1991), and these could also

ffect the cathodic reduction of O2 which involves proton trans-er. Previous reports for carbon-based electrodes found that for pH2–14 the reduction of O2 appears to be independent of pH, buts pH decreases the reduction becomes pH dependent (Taylor andumffray, 1975; Yang and McCreery, 2000). To test for the pH sen-

O2/Air) and mild hypoxia (100% N2/Air) on brain tissue O2 levels monitored usingedial prefrontal cortex and hippocampus of an anaesthetised rat. Filled gray areasr clarity) represent the SEM.

sitivity the buffer pH was changed from the standard physiological7.4 to 6.5 and 8.0. No change was observed in the O2 sensitiv-ity: −1.83 ± 0.14 nA/�M (7.4, n = 8); −1.64 ± 0.17 nA/�M (6.5, n = 4,P = 0.43); −1.91 ± 0.48 nA/�M (8.0, n = 4, P = 0.84). Similar findingshave previously been reported for O2 reduction at carbon fibre elec-trodes (CFEs) using fast cyclic voltammetry (FCV) (Zimmerman andWightman, 1991) and at multi-walled carbon nanotube modifiedglassy carbon electrodes using rotating disk electrode voltammetry(Kruusenberg et al., 2009).

3.4. Effect of ion changes

The media-dependence of redox reactions for electroactivespecies such as dopamine (DA), 3,4-dihydroxyphenylacetic acid(DOPAC), serotonin (5-HT) and 5-hydroxyindoleacetic acid (5-HIAA) have previously been investigated by several groups. Riceand co-workers found that the FCV sensitivities to DA and 5-HTat CFEs were 2–3-fold higher in non-physiological phosphate orHEPES-buffered saline compared to artificial cerebrospinal fluid(aCSF) which more accurately reflects the ionic composition of thebrain (Kume-Kick and Rice, 1998; Chen and Rice, 1999). Interest-ingly, the reverse was observed for the sensitivities of the acidmetabolites (DOPAC, 5-HIAA). Crespi reported shifts in oxidationpeaks and higher sensitivities using differential pulse voltamme-

try for aCSF that contained Ca2+ but not Mg2+, than in PBS thatcontained Mg2+ but not Ca2+ (Crespi, 1996). Such findings indicatethat electrode surface interactions differ for anions and cations andare consistent with adsorption or repulsion effects at the modified(Nafion® or electrically pre-treated) CFE surface.

140 F.B. Bolger et al. / Journal of Neuroscience Methods 195 (2011) 135–142

Fig. 4. Typical cyclic voltammograms recorded in vitro at 100 mV/s with a CPE inOmpo

c(iotropfiFi(ppfaa−it(P

3

s

Table 1In vitro CPA (−650 mV vs. SCE) response of carbon paste electrodes (n = 4) for a varietyof potential interferents at physiologically relevant concentrations.a Data expressedas a percentage of the O2 current at the reported extracellular level of ca. 50 �M(Zimmerman and Wightman, 1991).

Compoundb I (nA) O2 (%)

O2 (50 �M) −52.03 ± 2.43 100AA −0.12 ± 0.01 0.06 ± 0.03HVA −0.31 ± 0.01 0.06 ± 0.01l-Glutathione −0.22 ± 0.02 0.41 ± 0.02l-Cysteine −0.13 ± 0.02 0.25 ± 0.02Uric acid −0.13 ± 0.02 0.25 ± 0.025-HT −0.14 ± 0.02 0.27 ± 0.03l-Tryptophan −0.08 ± 0.02 0.16 ± 0.03DHAA −0.09 ± 0.03 0.16 ± 0.05l-Tyrosine −0.29 ± 0.05 0.55 ± 0.08DA −0.20 ± 0.07 0.37 ± 0.11DOPAC −0.04 ± 0.05 0.08 ± 0.105-HIAA −0.03 ± 0.06 0.06 ± 0.12

a 100 �M interferent or brain extracellular fluid concentration if known (O’Neill,1994; Yang et al., 1994; Brand et al., 1993): AA, 500 �M; HVA, 10 �M; l-glutathione,50 �M; l-cysteine, 50 �M; uric acid, 50 �M; 5-HT, 0.01 �M; dopamine, 0.05 �M;

2 (A) and N2 (A, inset) saturated artificial cerebrospinal fluid (aCSF). (B) Ampero-etric calibration plots for O2 (0–240 �M) at CPEs (n = 4) carried out using constant

otential amperometry at −650 mV (vs. SCE) in aCSF with and without either Ca2+

r Mg2+ ions.

As it has also been reported that the extracellular Ca2+ con-entration can fall under conditions of both electrical stimulationse.g. ca. 80 �M for 10 Hz at 10 s) (Kume-Kick and Rice, 1998) andntense tissue depolarisation (e.g. >1 mM for anoxic depolarisationr spreading depression) (Nicholson and Rice, 1988), we decidedo test the effect of ion changes (Ca2+ and Mg2+) on the CPE O2esponse in aCSF. No compensation was made for the decrease insmolality associated with removal of the ions as this change hasreviously been reported to be small (Chen and Rice, 1999). Werst confirmed the position for O2 reduction on the voltage axis.ig. 4A shows typical cyclic voltammograms recorded at 100 mV/sn N2 (background) and O2-saturated aCSF. As with PBS, −650 mVvs. SCE) is in the mass-transport limited region after the peakotential for O2 reduction, and all experiments in aCSF were thuserformed at this potential. O2 sensitivities for calibrations per-ormed in the range 0–240 �M O2 (Fig. 4B, n = 16) in the presencend absence of ions were similar (aCSF, −1.62 ± 0.11 nA/�M, n = 16;CSF (no Ca2+), −1.68 ± 0.14 nA/�M, n = 16, P = 0.73; aCSF (no Mg2+),1.50 ± 0.11 nA/�M, n = 16, P = 0.43) indicating that typical (small)

on changes in physiological media will not affect the CPE O2 reduc-ion signal. In fact, there was no significant difference between PBS−1.68 ± 0.09 nA/�M, n = 16) and aCSF (−1.62 ± 0.11 nA/�M, n = 16,= 0.67).

.5. Selectivity

The mammalian brain is a hostile environment for implantedensors as it contains electrode poisons (e.g. lipids and proteins)

DOPAC, 20 �M; 5-HIAA, 50 �M.b AA, ascorbic acid; HVA, homovanillic acid; 5-HT, 5-hydroxytryptamine; DHAA,

dehydroascorbic acid; DA, dopamine; DOPAC, 3,4-dihydroxyphenylacetic acid; 5-HIAA, 5-hydroxyindoleacetic acid.

and a large number of possible interfering species present at vary-ing concentrations ranging from low nM (e.g. catecholamines andtheir metabolites) to high �M (e.g. ascorbic acid, AA). The successor failure of a sensor designed for in vivo neurochemical appli-cations is generally decided by its ability to eliminate interferentsignals while maintaining sufficient sensitivity for its target ana-lyte (Phillips and Wightman, 2003). The selectivity of CPEs for O2reduction using CPA, relative to a variety of potential interferentspresent in the brain (Lowry et al., 1996), was thus characterizedin vitro. The compounds tested included the neurotransmitters DAand 5-HT, their metabolites DOPAC, homovanillic acid (HVA) and 5-HIAA and other electroactive species such as l-tyrosine, l-cysteine,l-tryptophan, l-glutathione, dehydroascorbic acid (DHAA) and thepurine metabolite uric acid (UA). While possible interferents foroxidised analytes (e.g. DA) have been well documented (Phillipsand Wightman, 2003; Lowry and O’Neill, 2005) this is not the casefor species such as O2 which are detected by electrochemical reduc-tion. The above list is thus not exhaustive but represents a rangeof standard interferents. It excludes other gaseous species (e.g. NO)which have standard potentials more positive than O2 as their effecton the current response has been reported to be negligible due totheir low (nM) concentrations (Falck, 1997).

The results are summarized in Table 1. Although slight positivechanges (<0.5 nA) were observed for most compounds follow-ing injection the signal tended to return towards baseline withinthe recording period (5 min) before the next injection. Also, suchchanges are negligible compared to the O2 sensitivity at physio-logical concentrations, and will thus have no effect on the CPE CPAresponse for O2, resulting in interference-free signals in vivo. Sim-ilar selectivity characteristics were found for O2 reduction at CPEsusing DPA (Lowry et al., 1996).

3.6. Stability

It is well accepted that noble metal (i.e. Clark-type) electrodesare susceptible to electrode poisoning in physiological media andgenerally require a protective membrane, which in some cases,

depending on the membrane, improves specificity (Clark, 1959;Pantano and Kuhr, 1995; Papkovsky, 2004; Hynes et al., 2006). CPEsin contrast are less prone to surface poisoning and are very sta-ble, even after weeks of continuous recording in the brain (Fillenzand O’Neill, 1986). As such, they do not require the use of pro-

F.B. Bolger et al. / Journal of Neuroscien

Fig. 5. Average (±SEM) baseline in vivo data (pooled from 12 animals—striatum, hip-puoA

tKpCebomOtWoaettrbotowcma

v(2c[2ra

4

−bttnlpta

ocampus, prefrontal cortex and nucleus accumbens) for CPEs (n = 17–26) recordedsing CPA at −650 mV over 21 days. Inset: average weekly baseline data recordedver the 3 week period and extended to 3 months using data from 3 animals (n = 12).ll data taken from a six-hour period covering morning/afternoon (10 am–4 pm).

ecting membranes. The basis of this stability was investigated byane and O’Neill (1998) who examined the effects of lipid androtein media on the electrooxidation of AA at both CPEs andFEs. CFEs were extensively poisoned whereas CPEs showed novidence of fouling. This novel resistance to poisoning appears toe due to the presence of the pasting silicone oil. Examinationf baseline in vivo data for CPEs implanted in freely moving ani-als suggests that this stability also extends to the reduction of

2 (Fig. 5): although there was a gradual decrease over the firsthree weeks (Week 1, −167.0 ± 1.5 nA; Week 2, −151.9 ± 2.3 nA;

eek 3, −141.7 ± 0.9 nA), no significant variation was observedver the 21-day period (P = 0.89, one-way ANOVA; n = 17–26, 12nimals). Using data from three animals (four implanted sensorsach) recorded out to 84 days indicates a stable baseline signal overwelve weeks (inset Fig. 5). It is important, however, to point outhat during a 24 h period the signal can exhibit numerous natu-ally occurring deviations from baseline levels. These changes cane rapid, occurring over periods ranging from seconds to minutes,r more prolonged, lasting one or more hours. The former tendso be associated with physiological phenomena such as naturallyccurring grooming and feeding, while the latter occurs mainlyith periods of intense activity typically observed during the dark

ycle (unpublished observations). Both are reflective of the fact theeasured real-time [O2] is the dynamic balance between supply

nd utilisation.It is noteworthy that such changes in signal (nA) can be con-

erted to units of pressure (mmHg), often used to represent pO2Offenhauser et al., 2005; Lecoq et al., 2009; Piilgaard and Lauritzen,009), using post in vivo calibration data and literature reportedoncentration and pressure data associated with the air-saturatedO2] at 37 ◦C (Forstner and Gnaiger, 2010): typical increases of2 nA observed for grooming, and 26 nA for feeding/drinking cor-espond to ca. 25 �M O2 and 29 �M O2 respectively, or 18 mmHgnd 24 mmHg.

. Conclusions

CPEs operating amperometrically at a constant potential of650 mV (vs. SCE) can be used reliably to monitor tissue O2 inrain extracellular fluid, and offer several advantages comparedo pulsed amperometric methods including simple experimen-al design, continuous real-time recording and low background

oise. In vitro and in vivo characterisation studies indicate a low

imit of detection, high sensitivity and interference free signals athysiological concentrations, rapid response time, no effect of pH,emperature, and ion changes, no dependence upon flow (stirring),nd long-term stability in vivo extending over weeks, with minimal

ce Methods 195 (2011) 135–142 141

consumption of O2. The significance of these results is highlightedby the recent finding that the tissue [O2] monitored using CPA ata carbon-based O2 sensor can serve as an index of changes in themagnitude of the blood oxygenation level dependent (BOLD) func-tional magnetic resonance imaging (fMRI) response (Lowry et al.,2010). The amperometric O2 signal thus provides a reliable awakeanimal surrogate of human fMRI experimentation, and is thus aneffective translational tool which can better enable the comparisonof pre-clinical and clinical research.

Acknowledgements

J.P.L. acknowledges the financial support of NUI Maynooth,Enterprise Ireland (CFTD/2008/107), Science Foundation Ireland(03/IN3/B376), and Eli Lilly and Co. Limited.

References

Amatore C, Szunerits S, Thouin L, Warkocz J-S. Mapping concentration pro-files within the diffusion layer of an electrode: part III. Steady-state andtime-dependent profiles via amperometric measurements with an ultramicro-electrode probe. Electrochem Commun 2000;2:353–8.

Bard AJ, Faulkner LR. Electrochemical methods: fundamentals and applications. NewYork: John Wiley & Sons, Inc; 1980a. p. 26–34.

Bard AJ, Faulkner LR. Electrochemical methods: fundamentals and applications. NewYork: John Wiley & Sons, Inc; 1980b. p. 19.

Baumgärtl H, Heinrich U, Lübbers DW. Oxygen supply of the blood-free perfusedguinea-pig brain in normo- and hypothermia measured by the local distributionof oxygen pressure. Pflügers Arch 1989;414:228–34.

Bazzu G, Puggioni GGM, Dedola S, Calia G, Rocchitta G, Migheli R, et al. Real-timemonitoring of brain tissue oxygen using a miniaturized biotelemetric deviceimplanted in freely moving rats. Anal Chem 2009;81:2235–41.

Bolger FB, Lowry JP. Brain tissue oxygen: in vivo monitoring with carbon pasteelectrodes. Sensors 2005;5:473–87.

Borland LM, Shi G, Yang H, Michael AC. Voltammetric study of extracellulardopamine near microdialysis probes acutely implanted in the striatum of theanesthetized rat. J Neurosci Methods 2005;146:149–59.

Bourdillon C, Thomas V, Thomas D. Electrochemical study of d-glucose oxidaseautoinactivation. Enzyme Microb Technol 1982;4:175–80.

Brand A, Richterlandsberg C, Leibfritz D. Multinuclear NMR-studies on the energy-metabolism of glial and neuronal cells. Dev Neurosci 1993;15:289–98.

Chen BT, Rice ME. Calibration factors for cationic and anionic neurochemicals atcarbon–fiber microelectrodes are oppositely affected by the presence of Ca2+

and Mg2+. Electroanalysis 1999;11:344–8.Clark Jr LC, Clark EW. Epicardial oxygen measured with a pyrolytic graphite elec-

trode. Al J Med Sci 1964;1:142–8.Clark Jr LC, Misrahy G, Fox RP. Chronically implanted polarographic electrodes. J

Appl Physiol 1958;13:85–91.Clark Jr LC, Wolf R, Granger D, Taylor Z. Continuous recording of blood oxygen

tensions by polarography. J Appl Physiol 1953;6:189–93.Clark Jr LC. Electrochemical device for chemical analysis. U.S. Patent 2913386; 1959.Cooper R. Local changes of intra-cerebral blood flow and oxygen in humans. Med

Electron Biol Eng 1963;1:529–36.Coplin WM, O’Keefe GE, Grady MS, Grant GA, March KS, Winn HR, et al. Accuracy

of continuous jugular bulb oximetry in the intensive care unit. Neurosurgery1998;42:533–9.

Crespi F. Carbon fibre micro-electrode and in vitro or in brain slices voltammetricmeasurement of ascorbate, catechol and indole oxidation signals: influence oftemperature and physiological media. Biosens Bioelectron 1996;11:743–9.

Deboer P, Damsma G, Fibiger HC, Timmerman W, Devries JB, Westerink BHC.Dopaminergic–cholinergic interactions in the striatum—the critical significanceof calcium concentrations in brain microdialysis. Naunyn Schmiedebergs ArchPharmacol 1990;342:528–34.

Duff A, O’Neill RD. Effect of probe size on brain extracellular uric acid monitoredwith carbon paste electrodes. J Neurochem 1994;62:1496–502.

Ewing AG, Dayton MA, Wightman RM. Pulse voltammetry with microvoltammetricelectrodes. Anal Chem 1981;53:1842–7.

El-Deab MS, Ohsaka T. Quasi-reversible two-electron reduction of oxygen at goldelectrodes modified with a self-assembled submonolayer of cysteine. Elec-trochem Commun 2003;5:214–9.

Falck D. Amperometric oxygen electrodes. Curr Separation 1997;16:19–22.Fillenz M, O’Neill RD. Effects of light reversal on the circadian pattern of motor-

activity and voltammetric signals recorded in rat forebrain. J Physiol (Lond)

1986;374:91–101.

Forstner H, Gnaiger E. Calculation of equilibrium oxygen concentration. In: GnaigerE, Forstner H, editors. Polarographic oxygen sensors. Aquatic and physiologicalapplications. Berlin/Heidelberg: Springer-Verlag; 2010. p. 321–36.

Foster TH, Hartley DF, Nichols MG, Hilf R. Fluence rate effects in photodynamictherapy of multicell tumor spheroids. Cancer Res 1993;53:1249–54.

1 oscien

G

G

H

H

H

H

J

J

J

K

K

K

K

K

L

L

L

L

L

L

L

L

M

M

M

M

toring: evaluation in the feline brain. Neurosurgery 1995;37:1168–76.

42 F.B. Bolger et al. / Journal of Neur

opinath SP, Valadka AB, Uzura M, Robertson CS. Comparison of jugular venousoxygen saturation and brain tissue pO2 as monitors of cerebral ischemia afterhead injury. Crit Care Med 1999;27:2337–45.

otoh F, Yoshiaki T, Meyer JS. Transport of gases through brain and their extravas-cular vasomotor action. Exp Neurol 1961;4:48–58.

itchman ML. Measurement of Dissolved Oxygen. New York: John Wiley & Sons,Inc; 1978. p. 103.

olmström N, Nilsson P, Carlsten J, Bowald S. Long-term in vivo experience of anelectrochemical sensor using the potential step technique for measurement ofmixed venous oxygen pressure. Biosens Bioelectron 1998;13:1287–95.

oward L, Gopinath SP, Uzura M, Valadka A, Robertson CS. Evaluation of a newfiberoptic catheter for monitoring jugular venous oxygen saturation. Neuro-surgery 1999;44:1280–5.

ynes J, Marroquin LD, Ogurtsov VI, Christiansen KN, Stevens GJ, Papkovsky DB, et al.Investigation of drug-induced mitochondrial toxicity using fluorescence-basedoxygen-sensitive probes. Toxicol Sci 2006;92:186–200.

eroschewski P, zur Linden D. A flow system for calibration of dissolved oxygensensors. Fresen J Anal Chem 1997;358:677–82.

in GY, Zhang YH, Cheng WX. Poly(p-aminobenzene sulfonic acid)-modified glassycarbon electrode for simultaneous detection of dopamine and ascorbic acid. SensActuators B: Chem 2005;107:528–34.

ustice Jr JB. Introduction to in vivo voltammetry. In: Justice Jr JB, editor. Voltammetryin the neurosciences: principles, methods, and applications. Clifton, NJ: HumanaPress; 1987. p. 45.

ane DA, O’Neill RD. Major differences in the behaviour of carbon paste and carbonfibre electrodes in a protein–lipid matrix: implications for voltammetry in vivo.Analyst 1998;123:2899–903.

han AS, Michael AC. Invasive consequences of using micro-electrodes and micro-dialysis probes in the brain. Trends Anal Chem 2003;22:503–8.

ayama T, Yoshimoto T, Fujimoto S, Sakurai Y. Intratumoral oxygen pressure inmalignant brain tumor. J Neurosurg 1991;74:66–9.

ruusenberg I, Alexeyeva N, Tammeveski K. The pH-dependence of oxygen reduc-tion on multi-walled carbon nanotube modified glassy carbon electrodes.Carbon 2009;47:651–8.

ume-Kick J, Rice ME. Dependence of dopamine calibration factors on media Ca2+

and Mg2+ at carbon-fiber microelectrodes used with fast-scan cyclic voltamme-try. J Neurosci Methods 1998;84:55–62.

ecoq J, Tiret P, Najac M, Shepherd GM, Greer CA, Charpak S. Odor-evoked oxygenconsumption by action potential and synaptic transmission in the olfactory bulb.J Neurosci 2009;29:1424–33.

owry JP, Boutelle MG, Fillenz M. Measurement of brain tissue oxygen at a carbonpaste electrode can serve as an index of increases in regional cerebral blood flow.J Neurosci Methods 1997;71:177–82.

owry JP, Boutelle MG, O’Neill RD, Fillenz M. Characterization of carbon paste elec-trodes in vitro for simultaneous amperometric measurement of changes inoxygen and ascorbic acid concentrations in vivo. Analyst 1996;121:761–6.

owry JP, Fillenz M. Real-time monitoring of brain energy metabolism in vivo usingmicroelectrochemical sensors: the effects of anesthesia. Bioelectrochemistry2001;54:39–47.

owry JP, Griffin K, McHugh SB, Lowe AS, Tricklebank M, Sibson NR. Real-timeelectrochemical monitoring of brain tissue oxygen: a surrogate for functionalmagnetic resonance imaging in rodents. NeuroImage 2010;52:549–55.

owry JP, Miele M, O’Neill RD, Boutelle MG, Fillenz M. An amperometric glucose-oxidase/poly(o-phenylenediamine) biosensor for monitoring brain extracellularglucose: in vivo characterisation in the striatum of freely moving rats. J NeurosciMethods 1998;79:65–74.

owry JP, O’Neill RD. Neuroanalytical chemistry in vivo using biosensors. In: GrimesCA, Dickey EC, editors. Encyclopedia of sensors. Stevenson Ranch, CA, USA:American Scientific Publishers; 2005. p. 501–24.

üebbers DW, Baumgäertl H. Heterogeneities and profiles of oxygen pressure inbrain and kidney as examples of the pO2 distribution in the living tissue. KidneyInt 1997;51:372–80.

a Y, Wu S. Simultaneous measurement of brain tissue oxygen partial pressure,temperature, and global oxygen consumption during hibernation, arousal, andeuthermy in non-sedated and non-anesthetized arctic ground squirrels. J Neu-rosci Methods 2008;174:237–44.

adsen PL, Linde R, Hasselbalch SG, Paulson OB, Lassen NA. Activation-inducedresetting of cerebral oxygen and glucose uptake in the rat. J Cereb Blood FlowMetab 1998;18:742–8.

artel D, Kuhn A. Electrocatalytic reduction of H2O2 at P2Mo18O626− modified glassy

carbon. Electrochim Acta 2000;45:1829–36.atsumoto H, Oda T, Hossain MA, Yoshimura N. Does the redox state of cytochrome

aa3 reflect brain energy level during hypoxia? Simultaneous measurementsby near infrared spectrophotometry and 31P nuclear magnetic resonance spec-troscopy. Anesth Analg 1996;83:513–8.

ce Methods 195 (2011) 135–142

McCreery DB, Agnew WF, Bullara LA, Yuen TG. Partial pressure of oxygen in brainand peripheral nerve during damaging electrical stimulation. J Biomed Eng1990;12:309–15.

Morton DB, Griffiths PHM. Guidelines on the recognition of pain and discom-fort in experimental animals and an hypothesis for assessment. Vet Rec1985;116:431–6.

Murr R, Berger S, Schuerer L, Peter K, Baethmann A. A novel, remote-controlledsuspension device for brain tissue pO2 measurements with multiwire surfaceelectrodes. Pflügers Arch 1994;426:348–50.

Nair PK, Buerk DG, Halsey Jr JH. Comparison of oxygen metabolism and tissue pO2

in cortex and hippocampus. Stroke 1987;18:616–22.Nicholson C, Rice ME. Use of ion-selective microelectrodes and voltammetric

microsensors to study the brain cell microenvironment. In: Boulton A, BakerGB, Walz W, editors. Neuronal microenvironment. Neuromethods. Clifton, NJ:Humana Press; 1988. p. 247–361.

O’Neill RD. Sensor–tissue interactions in neurochemical analysis with carbon pasteelectrodes in vivo. Analyst 1993;118:433–8.

O’Neill RD. Microvoltammetric techniques and sensors for monitoring neurochem-ical dynamics in vivo—a review. Analyst 1994;119:767–79.

O’Neill RD, Grunewald RA, Fillenz M, Albery WJ. Linear sweep voltammetry withcarbon paste electrodes in the rat striatum. Neuroscience 1982;7:1945–54.

Offenhauser N, Thomsen K, Caesar K, Lauritzen M. Activity-induced tissue oxygena-tion changes in rat cerebellar cortex: interplay of postsynaptic activation andblood flow. J Physiol (Lond) 2005;565:279–94.

Pantano P, Kuhr WG. Enzyme-modified microelectrodes for in vivo neurochemicalmeasurements. Electroanalysis 1995;7:405–16.

Papkovsky DB. Methods in optical oxygen sensing: protocols and critical analyses.In: Sen CK, Semenza GL, editors. Methods enzymol, 383; 2004. p. 715–35.

Phillips PEM, Wightman RM. Critical guidelines for validation of the selectivity ofin vivo chemical microsensors. Trend Anal Chem 2003;22:509–14.

Piilgaard H, Lauritzen M. Persistent increase in oxygen consumption and impairedneurovascular coupling after spreading depression in rat neocortex. J CerebBlood Flow Metab 2009;29:1517–27.

Rasmussen P, Dawson EA, Nybo L, van Lieshout JJ, Secher NH, Gjedde A. Capillary-oxygenation-level-dependent nearinfrared spectrometry in frontal lobe ofhumans. J Cereb Blood Flow Metab 2007;27:1082–93.

Revsbech NP. An oxygen microelectrode with a guard cathode. Limnol Oceanogr1989;34:472–6.

Rice ME, Nicholson C. Diffusion and ion shifts in the brain extracellular microen-vironment and their relevance for voltammetric measurements. In: Boulton A,Baker GB, Adams RN, editors. Voltammetric methods in brain systems. Neu-romethods. Clifton, NJ: Humana Press; 1995. p. 27–79.

Sam PM, Justice Jr JB. Effect of general microdialysis-induced depletion on extracel-lular dopamine. Anal Chem 1996;68:724–8.

Schneiderman G, Goldstick TK. Oxygen electrode design criteria and performancecharacteristics: recessed cathode. J Appl Physiol 1978;45:145–54.

Sharan M, Vovenko EP, Vadapalli A, Popel AS, Pittman RN. Experimental and theo-retical studies of oxygen gradients in rat pial microvessels. J Cereb Blood FlowMetab 2008;28:1597–604.

Silver IA. Some observations on the cerebral cortex with an ultra-micro, membranecovered, oxygen electrode. Med Electron Biol Eng 1965;3:377–87.

Taylor RJ, Humffray AA. Electrochemical studies on glassy carbon electrodes 2. Oxy-gen reduction in solutions of high pH (pH greater-than 10). J Electroanal Chem1975;64:63–84.

Thompson JK, Peterson MR, Freeman RD. Single-neuron activity and tissue oxygena-tion in the cerebral cortex. Science 2003;299:1070–2.

Travis RP, Clark Jr LC. Changes in evoked brain oxygen during sensory stimulationand conditioning. Electroencephalogr Clin Neurophysiol 1965;19:484–91.

Venton BJ, Michael DJ, Wightman RM. Correlation of local changes in extracellu-lar oxygen and pH that accompany dopaminergic terminal activity in the ratcaudate–putamen. J Neurochem 2003;84:373–81.

Yang CS, Chou ST, Lin NN, Liu L, Tsai PJ, Kuo JS, et al. Determination ofextracellular glutathione in rat-brain by microdialysis and high-performanceliquid-chromatography with fluorescence detection. J Chromatogr B: BiomedAppl 1994;661:231–5.

Yang HH, McCreery RL. Elucidation of the mechanism of dioxygen reduction onmetal-free carbon electrodes. J Electrochem Soc 2000;147:3420–8.

Zauner A, Bullock R, Di X, Young HF. Brain oxygen, CO2, pH, and temperature moni-

Zimmerman JB, Kennedy RT, Wightman RM. Evoked neuronal activity accompaniedby transmitter release increases oxygen concentration in rat striatum in vivo butnot in vitro. J Cereb Blood Flow Metab 1992;12:629–37.

Zimmerman JB, Wightman RM. Simultaneous electrochemical measurements ofoxygen and dopamine in vivo. Anal Chem 1991;63:24–8.